Increasing Value, Saving Lives: Health Care in a New Era - Keynote Address by Brent James
1 like650 views
The document discusses improvements in healthcare delivery through quality, utilization, and efficiency (QUE) studies across various clinical areas, highlighting the benefits of evidence-based protocols and data systems. It emphasizes the need for coordinated teams and reliable data tracking to enhance patient care and outcomes. Additionally, the presentation presents case studies demonstrating significant cost reductions and improved survival rates through structured clinical workflows.
Increasing Value, Saving Lives: Health Care in a New Era - Keynote Address by Brent James
- 1. Increasing Value, Saving Lives: Health Care in a New Era Brent C. James, M.D., M.Stat. Executive Director, Institute for Health Care Delivery Research Intermountain Healthcare Salt Lake City, Utah, USA Saskatchewan Health Quality Council Saskatchewan Health Care Quality Summit 2013 Evraz Place, Regina, Saskatchewan, Canada Thursday, 11 April 2013 -- 8:15a - 9:45a
- 2. Disclosures Neither I, Brent C. James, nor any family members, have any relevant financial relationships to be discussed, directly or indirectly, referred to or illustrated with or without recognition within the presentation. I have no financial relationships beyond my employment at Intermountain Healthcare.
- 3. Quality, Utilization, & Efficiency (QUE) Six clinical areas studied over 2 years: - transurethral prostatectomy (TURP) - open cholecystectomy - total hip arthroplasty - coronary artery bypass graft surgery (CABG) - permanent pacemaker implantation - community-acquired pneumonia pulled all patients treated over a defined time period across all Intermountain inpatient facilities - typically 1 year identified and staged (relative to changes in expected utilization) - severity of presenting primary condition - all comorbidities on admission - every complication - measures of long term outcomes compared physicians with meaningful # of cases (low volume physicians included in parallel analysis, as a group)
- 4. Intermountain TURP QUE Study Median Surgery Minutes vs Median Grams Tissue M L K J P B C O N A I D H E G F 0 20 40 60 80 100 0 20 40 60 80 100 Attending Physician Median surgical time Median grams tissue removed Gramstissue/Surgeryminutes
- 5. Intermountain TURP QUE Study 1500 1549 1568 1618 1543 1697 1913 2233 2140 2156 1598 1269 1164 1552 1556 1662 A B C D E F G H I J K L M N O P Attending Physician 0 500 1000 1500 2000 2500 Dollars 0 500 1000 1500 2000 2500 Average Hospital Cost
- 6. Total Hip Arthroplasty - LOS 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 0 2 4 6 8 10 12 14 16 LengthofStay(days) 1988 1989 1990 Month/Year 0 2 4 6 8 10 12 14 16 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 1988 1989 1990
- 7. Total Hip Arthroplasty - Cost 0 2 4 6 8 10 12 14 1988 1989 1990 0 2 4 6 8 10 12 14 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 1988 1989 1990 Month/Year Averagecostpercase($1,000s) 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3
- 8. Deming: Quality controls cost Quality Cost Forum internal internal Cost-benefit society - Waste: Savings Potential 25-40% > 50% (none) Inefficiency waste Quality waste
- 9. Deep post-op wound infections % prophylaxis given at optimal time 1985 1986 1991 40 58 96 % Infections 1.8 0.9 0.4 Est. decrease in infections relative to 1985 rate -- 33 51 Est. savings at $14,000 per case (in thousands) -- 462 714 National standard: 2 - 4% deep post-op wound infection rate LDSH Dept of Clinical Epidemiology
- 10. Deep post-op wound infections 1985 1994 38.0 37.1 40.0 99.1 19.0 5.3 43.0 14.3 % elective surgeries receiving prophylaxis % receiving first dose 0-2 hrs before incision % continuing prophylaxis 24 hrs after surgery Mean number of doses per case LDSH Dept of Clinical Epidemiology
- 11. NIH-funded randomized controlled trial assessing an "artifical lung" vs. standard ventilator management for acute respiratory distress syndrome (ARDS) discovered large variations in ventilator settings across and within expert pulmonologists created a protocol for ventilator settings in the control arm of the trial Dr. Alan Morris, LDS Hospital, 1991: We generalized the method
- 12. Problems with "best care" protocols Lack of evidence for best practice - Level 1, 2, or 3 evidence available only about 15-25% of the time Expert consensus is unreliable - experts can't accurately estimate rates using subjective recall (produce guesses that range from 0 to 100%, with no discernable pattern of response) - what you get depends on whom you invite (specialty level, individual level) Guidelines don't guide practice - systems that rely on human memory execute correctly ~50% of the time (McGlynn: 55% for adults, 46% for children) No two patients are the same; therefore, no guideline perfectly fits any patient (with very rare exception)
- 13. NIH-funded randomized controlled trial assessing an "artifical lung" vs. standard ventilator management for acute respiratory distress syndrome (ARDS) discovered large variations in ventilator settings across and within expert pulmonologists created a protocol for ventilator settings in the control arm of the trial Implemented the protocol using Lean principles (Womack et al., 1990 - The Machine That Changed the World) - built into clinical workflows - automatic unless modified - clinicians encouraged to vary based on patient need - variances and patient outcomes fed back in a Lean Learning Loop Dr. Alan Morris, LDS Hospital, 1991: We generalized the method
- 14. 1. Identify a high-priority clinical process (key process analysis) 2. Build an evidence-based best practice protocol (always imperfect: poor evidence, unreliable consensus) 3. Blend it into clinical workflow (= clinical decision support; don't rely on human memory; make "best care" the lowest energy state, default choice that happens automatically unless someone must modify) 4. Embed data systems to track (1) protocol variations and (2) short and long term patient results (intermediate and final clinical, cost, and satisfaction outcomes) 5. Demand that clinicians vary based on patient need 6. Feed those data back (variations, outcomes) in a Lean Learning Loop - constantly update and improve the protocol - provide true transparency to front-line clinicians - generate formal knowledge (peer-reviewed publications) Shared Baseline "Lean" protocols (bundles)
- 15. ARDS Protocol Compliance 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 0 10 20 30 40 50 60 70 80 90 100 ARDS Patient Number %ProtocolInstructionsFollowed 0 10 20 30 40 50 60 70 80 90 100 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59
- 16. Results: survival (for ECMO entry criteria patients) improved from 9.5% to 44% costs fell by ~25% (from $160k to $120k) physician time fell by ~50% Dr. Alan Morris, LDS Hospital, 1991
- 17. Key take-aways 1. No protocol perfectly fits any patient - solution: Shared Baseline "bundles" (mass customization = "patient centered care") 2. Serious limitations to protocol development - solution: a Learning System (embedded variance and outcomes tracking; continuous protocol review and tested improvement) 3. Reliance on human memory (craft of medicine) produces "55% execution" - solution: tools to embed protocols in workflows 4. Only two differences from traditional practice: It requires (1) coordinated teams with (2) reliable data systems
- 18. there is nothing new here ... It should have started in medicine ... except the idea that "it takes a team" (and true transparency = embedded data systems)
- 19. 07 Jan M ar M ay Jul Sep N ov08 Jan M ar M ay Jul Sep N ov09 Jan M ar M ay Jul Sep N ov10 Jan M ar Month 0 20 40 60 80 100 %compliance 0 20 40 60 80 100 ER bundle ICU bundle All components Sepsis bundle compliance
- 20. 04 Jan M ay Sep05 Jan M ay Sep06 Jan M ay Sep07 Jan M ay Sep08 Jan M ay Sep09 Jan M ay Sep10 Jan Month 0 0.1 0.2 0.3 0.4 0.5 Mortalityrate 0 0.1 0.2 0.3 0.4 0.5 Sepsis mortality - ER-ICU transfers 20.2% 8.0% ~116 fewer inpatient deaths per year 28 32 44 37 45 42 42 23 34 29 41 33 45 53 38 50 47 39 31 30 34 24 40 41 35 28 27 22 28 27 24 32 44 36 39 52 51 70 65 60 47 57 52 50 61 51 43 77 73 77 65 71 69 48 52 59 46 63 68 68 63 70 94 90 75 81 69 79 81 78 82 70 74 84 91n=
- 21. We count our successes in lives ... Lesson 1
- 22. 6.66 3.36 2.47 2.65 3.44 4.26 37 38 39 40 41 42 Weeks gestation 0 2 4 6 8 10 PercentNICUadmissions 0 2 4 6 8 10 Deliveries w/o Complications, 2002 - 2003 8,001 18,988 33,185 19,601 4,505 258n = NICU admits by weeks gestation
- 23. Elective inductions < 39 weeks 5.5 5.1 6.66.3 6 5.3 8.2 5.45.7 6.66.6 7.9 6.4 7.67.6 4.6 3.5 4.54.3 6.5 3.2 2.62.3 4.2 2.1 3.23.4 2.4 5 3 3.5 26.726.9 2929.2 25.3 27.6 20.4 19.1 16.5 15.2 8.4 10.7 8.1 6.8 5.96.1 6 5.1 6.3 Jan 01 M ar M ay Jul Sep N ovJan 02 M ar M ay Jul Jan 03 M ar M ay Jul Sep N ovJan 04 M ar M ay Jul Sep N ovJan 05 M ar M ay Jul 0 5 10 15 20 25 30 %electiveinductions<39weeks 0 5 10 15 20 25 30 382 372 490 415 430 435 422 455 430 382 356 337 372 366 455 n = 423 453 473 476 512 475 602 557 667 564 637 578 541 573 533 505 501 474 536 562 545 535 493 520 494 430 440 500 421 474 562 549 555 528 491
- 24. 33 31.4 36.1 28.3 17.7 15.1 17.6 14.4 14.3 5.8 4.5 2.1 0 20 8.2 8.5 3.6 3.4 3.9 3.2 2.4 1.1 0.9 1 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Bishop score 0 5 10 15 20 25 30 35 40 Percentc-sections 0 5 10 15 20 25 30 35 40 Unplanned c-section rates Electively induced patients by Bishop score, Jan 2002 - Aug 2003 10 49 130 274 567 856 1114 1266 1062 737 415 86 19 18 35 61 99 164 278 375 487 453 346 179 47 7 Multips Primips n
- 25. 22.1 20.7 17.4 15.7 15 13.8 12.6 11.6 10.4 9 9 7.5 8.2 12.4 12 10.8 10.1 9.2 8.1 7.6 7.1 6.4 5.9 5.5 5.1 4.1 1 2 3 4 5 6 7 8 9 10 11 12 13 Bishop score 0 5 10 15 20 25 Hours 0 5 10 15 20 25 Average hours in labor & delivery Electively induced patients by Bishop score, Jan 2002 - Aug 2003 10 49 130 274 567 856 1114 1266 1062 737 415 86 19 18 35 61 99 164 278 375 487 453 346 179 47 7 Multips Primips n
- 26. 15.3 14 15.3 14.514.7 11.6 12.8 11.8 12.612.8 15.1 12.1 9.9 8.8 6.8 6.5 6 6.1 7.6 6.5 6.6 5.2 4.9 8.4 4.3 4.3 4.5 6.1 5.4 4.4 3.9 53 53 63 53 57 45 56 52 41 52 62 46 49 35 21 21 26 28 34 28 22 18 20 35 15 18 15 18 25 21 20 110 87 119 109 124 91 107 94 100 105 118 87 81 67 57 57 46 52 60 55 49 37 33 67 30 30 36 48 45 37 34 Jan 2003 Feb M ar A prM ay Jun JulA ug Sep O ctN ov D ec Jan 2004 Feb M ar A prM ay Jun JulA ug Sep O ctN ov D ec Jan 2005 Feb M ar A prM ay Jun Jul 0 20 40 60 80 100 120 140 Numberofpatients 0 10 20 30 40 50 %ofallprimiparousdeliveries Primiparous elective inductions Bishop's score < 10 Bishop's score < 8 Goal: Reduce "inappropriate" nullip inductions by 50%
- 27. Elective induction: length of laborJan 2001 M ar M ay Jul Sep N ov Jan 2002 M ar M ay Jul Sep N ov Jan 2003 M ar M ay Jul Sep N ov Jan 2004 M ar M ay Jul Sep N ov Jan 2005 M ar M ay Jul Sep N ov 0 2 4 6 8 10 Averagehoursfromadmissiontodelivery 0 2 4 6 8 10 8.5 7.9 7.5 7.1 6.9 (note: includes all elective inductions)
- 28. Overall c-section rate 96 97 98 99 2000 01 02 03 04 05 06 0% 10% 20% 30% 40% Percentc-sectionsoverall 0% 10% 20% 30% 40% National Intermountain
- 29. 2001 2002 2003 2004 0 2,000,000 4,000,000 6,000,000 8,000,000 10,000,000 Coststructureimprovement($) 0 1,000,000 2,000,000 3,000,000 4,000,000 5,000,000 6,000,000 7,000,000 8,000,000 9,000,000 10,000,000 Cumulativeannualtotal($) Combined maternal and neonatal variable cost Deliveries without complications resulting in normal newborns Actual - expected cost, based on year-end 2000 with PPI inflation Quality-based cost improvement
- 30. Very often, better care is cheaper care ... Lesson 2
- 31. 50+% of all resource expenditures in hospitals is quality-associated waste: recovering from preventable foul-ups building unusable products providing unnecessary treatments simple inefficiency Andersen, C. 1991 James BC et al., 2006
- 32. No good deed goes unpunished Neonates > 33 weeks gestational age who develop respiratory distress syndrome Treat at birth hospital with nasal CPAP (prevents alveolar collapse), oxygen, +/- surfactant Transport to NICU declines from 78% to 18%. Financial impact (NOI; ~110 patients per year; raw $): Birth hospital Transport (staff only) Tertiary (NICU) hospital Delivery system total Integrated health plan Medicaid Other commerical payers Payer total Before 84,244 22,199 958,467 1,064,910 900,599 652,103 429,101 1,981,803 After 553,479 - 27,222 209,829 736,086 512,120 373,735 223,215 1,109,070 Net 469,235 - 49,421 -748,638 -328,824 388,479 278,368 205,886 872,733
- 33. Current U.S. payment mechanisms Actively incent overutilization: do more, get paid more - even when there is no health benefit I am paid to harm my patients (paid more for complications) Actively disincents innovation that reduces costs through better quality (a key success factor for the rest of the U.S. economy) Very strong, deep, wide evidence showing exactly this effect throughout U.S. healthcare
- 34. 1. ACOs, AMHs, bundled payment, shared savings, pay for value: sophisticated forms of capitation - provider at (financial) risk ... but with far better data systems for (1) quality measurement and (2) risk adjustment 2. Represent "managed care at the bedside" - ask clinical teams at the bedside to manage the care, not distant and disengaged insurance companies 3. More than 80% of cost saving opportunities live on the clinical side; 70+% of clinical improvement activities reduce costs by freeing up care delivery capacity (technically, "fixed cost leverage"). Capitation makes a comeback
- 35. A fundamental shift in focus The past: 1. "Top-line" revenue enhancement - Systems designed around documentation to support FFS payment, clinical decision support as a secondary "bolt-on" 2. Quality defined as regulatory compliance - e.g. - CMS Core Measures - Pay for Value - Meaningful Use The future: 1. Quality becomes the core business - Demonstrated performance for key clinical processes - Systems designed around clinical decision support (process management), producing documentation as an integrated by-product 2. "Bottom-line" cost control and waste elimination in a "provider at risk" financial environment
- 36. 1. Quality improvement is the science of process management 2. A focus on process management forces patient-centered care - care built along the full continuum of care; not buildings, technologies, or physicians 3. Combining patient- centered care with various levels of provider-at-financial-risk forces population-level care and the triple aim - collaborating with other community organizations (churches, schools, local governments, etc.) to promote best health and high function Getting to the Triple Aim
- 37. Better has no limit ... an old Yiddish proverb